AI NEWS PAGE DOSSIER
Transformative-AI forecasts, METR, technical safety, and years spent trying to put numbers on loss-of-control risk.
Cotra came out of Berkeley's effective-altruism scene, interned at GiveWell, then became Open Philanthropy's first research hire dedicated specifically to the new organization. She built one of the best-known quantitative forecasts for transformative AI, shifted into technical alignment and grantmaking as her timelines shortened, and ultimately moved into hands-on loss-of-control evaluation at METR.
2015
While studying electrical engineering and computer science at UC Berkeley, Cotra co-ran an effective-altruism course and interned with GiveWell. She later said the student work helped put her on GiveWell and Open Philanthropy's radar.
RECEIPTS: Effective Altruism ForumreportedAjeya Cotrafirst-party
Jul 2016
After graduating early from Berkeley, Cotra joined Open Philanthropy full time. She later described herself as the first research employee hired specifically for Open Phil rather than for the combined GiveWell/Open Phil operation.
RECEIPTS: Effective Altruism ForumreportedAjeya Cotrafirst-party
Sep 18, 2020
Cotra released her draft biological-anchors report, a quantitative framework for estimating when the compute required for transformative AI might become affordable. The project became one of the best-known attempts to put explicit probabilities on advanced-AI timelines.
RECEIPTS: Ajeya Cotrafirst-party
2022
Cotra published both an updated view of her AI timelines and a long argument that, without specific countermeasures, the easiest path to transformative AI could lead to AI takeover. Forecasting and alignment risk had become one problem in her work.
RECEIPTS: Ajeya Cotrafirst-partyAjeya Cotrafirst-party
Mar 26, 2023
Cotra launched Planned Obsolescence with Kelsey Piper while funding AI-alignment research at Open Philanthropy. The project was explicitly aimed at thinking ahead to a world in which AI systems could be making many of society's most important decisions.
RECEIPTS: Ajeya Cotrafirst-party
2024
Cotra led the technical AI safety program at what is now Coefficient Giving, shifting from individual forecasting toward deciding which research programs and grants might reduce risks from more capable AI systems.
RECEIPTS: METRreportedCoefficient Givingreported
2026
Cotra joined METR's technical staff to work on threat modeling and loss-of-control risk assessment for advanced AI, moving from funding and forecasting safety work into directly evaluating evidence about frontier systems.
RECEIPTS: METRreported